NGINX Service
- Alerts: https://alerts.gitlab.net/#/alerts?filter=%7Btype%3D%22nginx%22%2C%20tier%3D%22sv%22%7D
- Label: gitlab-com/gl-infra/production~“Service::NGINX”
Logging
Section titled “Logging”Troubleshooting Pointers
Section titled “Troubleshooting Pointers”- Disk space alerts (production)
- CustomersDot main troubleshoot documentation
- design.gitlab.com Runbook
- HAProxy Management at GitLab
- GitLab Query Language (GLQL)
- Service Error Rate
- Service Operation Rate
- Filesystem errors are reported in LOG files
- monitor.gitlab.net not accessible or return 5xx errors
- Gitlab.com on Kubernetes
- Packagecloud Infrastructure and Backups
- Life of a Git Request
- Life of a Web Request
- Managing Chef
- Google mtail for prometheus metrics
- ../uncategorized/namespace-restore.md
- version.gitlab.com Runbook
- Nginx is down
Summary
Section titled “Summary”NGINX sits in front of our Puma services. It provides a bit of protection between end users and puma workers to prevent saturation of threads.
Architecture
Section titled “Architecture”For Virtual Machines, this is deployed via our Omnibus package and runs as a service that recieves traffic.
For Kubernetes, this is deployed using the NGINX Ingress controller managed by our helm chart. https://docs.gitlab.com/charts/charts/nginx/